INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Sarcoidosis is a systemic disease characterized by granuloma formation and lymphocyte accumulation in the involved organs, mainly in the lung. Bronchoalveolar lavage (BAL) is broadly used to assess inflammatory processes in the alveoli. Absolute numbers of cells in BAL are increased resulting in a predominance of T lymphocytes in the differential cell counts. These T lymphocytes are primarily of the CD4⁺ phenotype; morphologically they appear activated, and express cell surface markers such as interleukin 2 receptors (IL-2R) and class II major histocompatibility complex (MHC II) molecules, characteristic of T cells recently stimulated via the T cell antigen receptor (TCR) (1). This, and additional findings such as decreased surface density of TCR and the production of IL-2 and IFN- by BAL cells in sarcoidosis (2), suggest the participation of an antigen in the immune-mediated granuloma formation in sarcoidosis. In general, antigens are digested, degraded into peptide fragments by the antigen presenting cells, and then presented on the surface membrane in context with MHC molecules. T cells specifically recognize processed antigens in conjunction with MHC molecules through their T cell receptors (TCR). More than 95% of the TCR's consist of an  and  chain heterodimer, the remaining 5% are characterized by a  and  chain heterodimer. TCR  chains consist of variable (V), diversity (D), joining (J), and constant (C) regions, whereas the  chain consists only of the V, J, and C regions. The TCR's antigen specificity is defined by the V domains encoded by variable and (diversity) joining gene elements, which are rearranged during the T cell differentiation. Forty-seven V genes can be grouped into 24 families defined by their sharing of more than 75% amino acid sequence homology (5). This region and a similar region in the -chain are thought to play crucial roles in defining T cell clonal specificity by coding for amino acids that interact with specific peptide-MHC ligands (6).

Antigen recognition by the T cell and their subsequent activation results in cytokine release and proliferation of the T cells with the appropriate TCR. The proliferation of the activated T cells can lead to an increase in the proportion of the respective V families. Conversely, shifts in the frequencies of the V families suggests a proliferative process driven by a either nominal antigen or a superantigen. In the last few years several authors have searched for these disparities in the V-repertoire of the sarcoid patients (1, 7). The results of these studies, however, disclosed a high heterogeneity of the V-repertoire of sarcoid patients. Moller and colleagues showed an increase of the V8 family in the BAL of sarcoid patients (7), which could not be established by others (11, 12). Bellocq and coworkers found a normal distribution of the V family except V19 (11), others found an increase in V2.3 (9, 12) or C1 (8). Interestingly, Klein and colleagues could demonstrate that in intradermal Kveim-Siltzbach reaction sites, the proportions of V2, V3, V5, V6, and V8 are increased compared to that of blood (10). One problem with these studies is that most of them compare BAL-T cells with T cells from the peripheral blood (PB). Although there are signs of inflammation in the alveoli that make the use of BAL cells useful, the compartment where the disease is located is the interstitium, which is not necessarily probed by the BAL. There is a dramatic increase in the percentage of CD4 positive T cells in the BAL of sarcoid patients. Therefore, it is possible that shifts in the V family repertoire differ in the two CD4 and CD8 defined T cell subpopulations in the compartments blood, interstitium and BAL. However, this cannot be probed by current PCR-methods. We therefore cloned T cells from the three compartments and analyzed their TCR-V usage and their CD4 and CD8 phenotype.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

We cloned cells from 16 patients with pulmonary sarcoidosis. From an additional 31 patients, we tried to establish T-cell clones; however, in these cases we failed to succeed in cloning from BAL as well as from TBB. The diagnosis of sarcoidosis was established using defined criteria, including transbronchial biopsy (TBB). At the time of investigation, the patients were not receiving therapy. One patient with tuberculosis and one patient with hypersensitivity alveolitis (Aspergillus fumigatus) were used for comparison.

Reagents, Media, and Equipment Applied

For the purpose of this investigation we have applied the following media and reagents: RPMI 1640 medium (Seromed, Berlin, Germany); PBS stock solution (GibcoBRL, Paisley, UK); AB-medium: RPMI 1640 supplemented with 1% glutamin (Gibco), 1% HEPES (Gibco), 1% sodium pyruvate (Gibco), 1% penicillin-streptomycin solution (Gibco), 2% inactivated human serum (Blood Bank, Mainz, Germany); cloning medium: AB-medium conditioned for 24 h with 10⁶ peripheral blood mononuclear cells of healthy donors per ml activated with 2 µg/ml PHA (Wellcome, Hannover, Germany) diluted 1:1 with fresh AB-medium plus 20 ng/ml of human recombinant IL-2 (Cetus, Emeryville, CA); feeder cells: irradiated (40 Gy) peripheral blood mononuclear cells isolated from buffy-coat of healthy donors and stored in aliquots of 10⁷ cells in liquid N₂; collagenase, DNAse, elastase (Sigma, München, Germany); anti-CD3 (OKT3; Ortho, Neckargmünd, Germany); anti-CD4 (OKT4; Ortho); anti-CD8 (OKT8; Ortho); anti-T-cell-receptor / (BW242/412; Behring, Marburg, Germany); rabbit anti-mouse IgG (P161; DAKO, Glostrup, Denmark); chromogen solution: 4 µg 3-amino-9-ethylcarbazol (AEC, Sigma) dissolved in 1 ml dimethyl formamide (Sigma) and made up to 15 ml with 0.1 M sodium acetate buffer (pH 4.9); 96-well culture plates (Nunc, Wiesbaden, Germany); 60-well microplates (Terasaki) (Nunc); glutaraldehyde (Merck, Darmstadt, Germany). Antibodies to V8 (MX-6) were a gift, courtesy of Dr. S. Carell (Ludwig Institute for Cancer Research, Epalinges, Switzerland), V2, V3, V13.3, V17, and V19 were from Dianova (Hamburg, Germany), V5.2-3, V6.7, and V12.1 from T Cell Diagnostics Inc. (Cambridge, MA). The normal PBL staining pattern of these antibodies was taken from the suppliers. The A. fumigatus-antigen was purchased from HAL Allergens (Düsseldorf, Germany).

Preparation of Bronchoalveolar, Lung Biopsy, and Blood Mononuclear Cells

BAL was performed according to standard protocols. In brief, 200 to 300 ml sterile saline (0.9% NaCl) were instilled in 25 ml aliquots into a lingula or middle lobe segment. Every aliquot was immediately aspirated. Recovered BAL fluid was filtrated and centrifuged at 500 × g. The cell pellet was resuspended and the cells were washed three times with RPMI 1640. PB was obtained by venipuncture, and mononuclear cells were isolated by Ficoll Hypaque gradient centrifugation. Lymphocytes of pulmonary parenchyma were obtained by enzymatic digestion of tissue samples taken by TBB from subpleural regions of the lung as described (13). In brief, the specimens were washed in PBS and incubated at 37° C in RPMI 1640 with collagenase (150 U/ml), elastase (10 U/ml), and deoxyribonuclease (50 U/ml). After digestion, the cells were washed and resuspended in culture medium.

Cell Cloning

Cells from the three compartments tested were resuspended in AB-medium supplemented with 40 U/ml IL-2 at a density of 1 × 10⁶ cells/ ml and cultivated in 96-well plates coated with OKT3 (5 µg/ml). After 3-7 d blasts were collected, seeded at a cell density of 0.5 to 2 cells/ well with feeder cells (1-2 × 10⁵/ml) in cell cloning medium, and incubated at 37° C in a humidified atmosphere with 5% CO₂. Restimulation was performed every 10-14 d with fresh cloning medium and 1-2 × 10⁵/ml feeder cells. Clones were expanded to 10⁶ cells and subsequently phenotyped.

Clone Phenotyping

The surface expression of CD3, CD4, CD8, TCR///, and the V-phenotype were tested by a cell-ELISA in micro-well plates. Each plate was incubated for 1 h with 10 ml PBS-diluted poly-L-lysin solution (1:20) at room temperature (RT) and then rinsed with PBS. The cells were washed, resuspended in PBS, and a 10-µl aliquot with approximately 10⁴ cells was added to each well. After 30-min sedimentation, the PBS supernatant was discarded and the cells were incubated for 1 h with 10 µl of the appropriate antibody solution/well. The supernatants were discarded, cells were fixed for 4 min at RT with 0.01% glutar-aldehyde and intensively washed. Incubation with peroxidase-labeled rabbit anti-mouse antibodies was performed for 30 min at RT. After washing, the plate was covered with chromogen solution for 15 min. The plates were then rinsed and analyzed with an inverted microscope for presence of red-colored cells indicating the binding of the particular antibody by an individual clone. Clonality was considered when the cells were stained by only one antibody and the staining of the cells was > 95%.

Statistics

Results are given as mean ± SD. Comparisons of the CD4/CD8 ratio before and after the cloning procedure were done by the Wilcoxon signed rank test and correlations were tested by estimation of the Spearman Rank Correlation Coefficient (). The significance of the positive stained clones versus the expected values or differences between the compartments were estimated by the chi-squared test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

General Approach

Among several tested cloning protocols, the one described previously (preactivation of the T cells with -CD3 monoclonal antibody and subsequent cloning with PHA and feeder cells) was the only protocol able to elicit clones from each of the compartments PB, BAL, and lung interstitium with sufficient efficiency. In total, 1,359 clones from all compartments of 16 patients with sarcoidosis were generated. The majority of clones were derived from the PB (n = 636), followed by the biopsies (n = 372), and the BAL (n = 351). In 31 patients, however, we failed to generate clones from BAL as well as TBB and in some cases in PBL (5/31). Interestingly, among these cases the percentage of patients with spontaneous remission within the following 6 months was greater (15/31 patients, 48%) than among the patients in which cloning was successful (2/16 patients, 12%; p < 0.03 [chi-squared test]).

CD4/CD8 Ratio of the Clones

The cloning procedure increases the CD4/CD8 ratio of the clones compared with the original cell populations (p < 0.01). However, correlating the CD4/CD8 ratios of the original cell population with those of the clones from PBL and BAL revealed a significant positive correlation ( = 0.7, p < 0.05; Figure 1).

View larger version (15K): [in this window] [in a new window]   Figure 1.   Correlation of the CD4/CD8 ratio of the clones and the original cell population. CD4/CD8 ratios of the clones and the original cell population disclose a significant positive correlation ( = 0.7, p < 0.05; Spearman Rank Correlation Coefficient).

1,277 clones from all compartments were found to be CD3 positive (blood: 581; TBB: 364; BAL: 332), and only these clones were analyzed further. The overall CD4/CD8 ratio was 842:435, yielding a quotient of 1.94. However, there were significant differences in the CD4/CD8 ratio of the clones achieved from the different compartments. Clones from the PB displayed a CD4/CD8 ratio of 2.3 ± 2.1 (n = 14, range from 0.1 to 6.3) and from TBB 2.7 ± 3.2 (n = 9, range from 0.02 to 8.4; Figure 2). Clones from BAL exhibited a CD4/CD8 ratio of 9.6 ± 3.6 (n = 7, range from 1.3 to 29; Figure 2). There was no statistical difference between the CD4/CD8 ratio of the clones from the PB and the TBB (p > 0.5) whereas both ratios were statistically distinct from the CD4/CD8 ratio of the BAL (p < 0.03 in both cases). However, correlating CD4/CD8 ratios of BAL and TBB revealed a striking positive correlation between both compartments ( = 1, p < 0.05), indicating that the increase of the CD4/CD8 ratio, although lower in the TBB, is concordant in both compartments. No correlation of the CD4/CD8 ratio of the PB to that of the BAL or TBB could be observed.

View larger version (16K): [in this window] [in a new window]   Figure 2.   CD4/CD8 ratios of the clones from the compartments bronchoalveolar lavage (BAL, n = 332), transbronchial biopsy (TBB, n = 364), and peripheral blood (PB, n = 581). The rectangle gives the range from the 25 to the 75% percentile, the horizontal line indicates the median, and the vertical line indicates the range from the 10 to the 90% percentile. Data beyond this range are individually depicted (circles). The CD4/CD8 ratio of the clones from BAL differ significantly from PB or TBB (p < 0.03, respectively).

Distribution of the V Families among the Clones from the Three Compartments

Peripheral blood. The most frequent V family in clones from the PB of the patients was V5 (11.8%), which was more dominant in the CD4⁺ population (13.7%) than in the CD8⁺ population (5.7%; Table 1). Compared with the expected values (1-7%) this family was overexpressed in the total CD3⁺ T cells as well as in the CD4⁺ subpopulation (p < 0.02, respectively) but not in the CD8⁺ subpopulation. The percentage of V19⁺ clones (7.2% in all clones, 8.9% in the CD4⁺ population) exceeded the expected range for this family; however, this increase did not reach statistical significance. Staining for the other seven V families revealed values within the expected range for the respective V family.

Bronchoalveolar lavage. In the BAL, V5 exceeded the expected value in CD3⁺ clones (28.6%, p < 0.02) and in the CD4⁺ subpopulation (27.2%, p < 0.001; Table 1). Due to the low number of clones (n = 15), the dramatic increase in V5 positive clones (53.3%) did not reach significance in the CD8⁺ subpopulation. V8 positive clones had also increased in the CD3⁺ clones (14.3%, p < 0.05) and in the CD4⁺ subpopulation (15.1%, p < 0.03) but not in the CD8⁺ subpopulation (0%). The percentages of V12, V13.3, and V19 were also increased but did not reach significance (Table 1).

Transbronchial biopsies. In TBB an increase could be demonstrated in the V5, V6, V13.3, and V19 families in CD3⁺ clones and in the CD4⁺ subpopulation. The increase of the V5 percentage in the TBB almost reached a significant level (p = 0.06); however, the increase of the other families did not reach statistical significance (Table 1).

Each of the employed antibodies stained at minimum for one of the tested clones except the antibody for V17, which stained none of 1,277 tested clones.

Comparing the V-distribution of the different compartments revealed a significantly higher usage of the TCR V5 in the BAL than in the PB or the TBB (p < 0.001). This pattern was observable in the total CD3⁺ clones as well as in the CD4⁺ and the CD8⁺ subpopulations. A similar distribution could be demonstrated for the V8 usage. Again, this family was more frequently represented in the BAL than in the PB or the TBB. However, in this case, this pattern could only be found in the entire CD3⁺ population or in the CD4⁺ subpopulation and not in the CD8⁺ subpopulation. Further disparities in the usage of other V's could not be detected.

Compartmentalized V Usage in Individual Patients

Although the overall distribution of the V families among the clones, with the exception of V5, did not show significant changes from the expected, striking differences could be observed between different compartments of patients. In patient #115 V5 and V8 were equally distributed whereas V6 and V13 could only be demonstrated in PB but not in the corresponding biopsy. However, the frequency of V19 was considerably higher in the TBB than in the PB (Figure 3). In patient #119, the opposite was true for this V family; a high frequency in the TBB and a normal percentage in the PB. V2 and V3 could only be demonstrated in the PB but not in the TBB (Figure 3). High frequencies for V2, V5, and V8 were found in the TBB of patient #121, which were absent or within the normal range in the corresponding PB. The frequencies for V6, V13, and V19 were in the normal range in the PB and the TBB. V6 and V19 were also missing in the TBB but normally expressed in the PB (Figure 3). In the TBB of patient #122, V5 and V8 were both found within the normal range and no staining for the other families could be observed. In the PB of this patient, high frequencies of V2, V5, and V19 were found, whereas V8 was found in the expected range and no staining for the rest of the families could be observed. However, high frequencies of V5, V8, V12, and V13 without evidence of any other V family could be detected in BAL (Figure 3).

View larger version (33K): [in this window] [in a new window]   Figure 3.   V family percentages of the clones from different compartments of sarcoid patients. T cells from different compartments were cloned and the V-usage was determined with monoclonal antibodies. In the upper two rows two compartments are investigated (PB and either TBB or BAL), in the lower row all three compartments are investigated. At least 20 clones per individuum were analyzed.

In order to compare our findings from the sarcoid patients, we also cloned T cells from one patient with tuberculosis and one patient with alveolitis and sensitivity to A. fumigatus. In the TBB of the tuberculoid patient only V5 and V8 could be detected and both families showed an increased frequency. In the PB of this patient, V5 had also increased; whereas V8 showed a normal percentage. The relative occurrence of V19 was increased in the PB but this family was missing in the TBB. In the hypersensitivity alveolitis patient, clones generated with -CD3 differed in their V-composition from the clones elicited with Aspergillus antigen. Aspergillus-induced-clones from the BAL bore V6, V8, and V19 whereas clones from the same compartment pre-stimulated with -CD3 used V2, V8, and V19. Aspergillus clones from the PB expressed V3, V5, V6, V8, and V19; whereas PB clones with -CD3 utilized V3, V8, and V19 (Figure 4).

View larger version (17K): [in this window] [in a new window]   Figure 4.   (A) V usage of clones from a patient with tuberculosis. T cells from PB and TBB were cloned and the V-usage was determined with monoclonal antibodies. (B) and (C ) show T cell clones from a patient sensitized with Aspergillus fumigatus. The T cells were prestimulated either with A. fumigatus antigen (B) or with anti-CD3 (C ), cloned, and the V-usage was determined with monoclonal antibodies.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mononuclear cell alveolitis and increase of the CD4/CD8 ratio are a well known phenomena accompanying sarcoidosis. We report here that (1) there is a high variability in the V-usage in the three compartments of the body of individual patients with sarcoidosis; (2) With the exception of V5 there is no predominant V family in the average; (3) The most prominent changes in the V family distribution could be demonstrated in the BAL; (4) CD4/CD8 ratio in the interstitium also increases but does not reach the level of that of the BAL.

To support the clinical diagnosis of sarcoidosis and to evaluate the activity of the disease, BAL is widely used to assess the cell composition of the alveoli. However, sarcoidosis is an interstitial lung disease, and there is an ongoing discussion as to whether or not BAL reflects the processes in the interstitium (14). Although the increase of the CD4/CD8 ratio in the BAL is well known, it is not yet clear whether there is a comparable increase in the interstitium. Additionally, in the past few years, several authors have pointed out that the T cells in the BAL are oligoclonal, implicating an antigen-driven influx or proliferation of the T cells in the alveoli. Hence, we cloned T cells from the three compartments PB, BAL, and interstitium and compared the TCR-V-usage and the clonal CD4/ CD8 ratios in these different compartments.

It was not difficult to obtain clones from the PB by direct cloning from purified T cells (E-Rosetting) with PHA; however, we failed to generate clones from nonpreactivated BAL-cells. T cells from the interstitium, isolated by enzymatic digestion of pieces of TBB, were very low in numbers and were also refractory in their response to PHA. The phenomenon that T cells from the BAL of either normal subjects or patients with sarcoidosis are refractory to proliferative signals is well known (15) and is also reported in hypersensitivity pneumonitis (16). Prestimulation of the cells with immobilized -CD3 in IL-2 containing medium overcame the refractory state in most cases, and the activated cells could be restimulated and expanded by repeated stimulation with PHA and heterologous feeder cells. However, even when the cells were prestimulated in various cases we failed to generate clones from either one or two compartments. In a number of cases even -CD3 stimulation failed to stimulate the T cells. Interestingly, in these cases, approximately one half of the patients had a spontaneous remission within the following 6 mo.

Recently, we have been able to demonstrate that BAL cells from patients with spontaneous remission release more TGF than BAL cells from either control subjects or from sarcoid patients with indications for therapy (17). TGF inhibits IL-2 mediated effects on T cells (18). Although we have no proof, it is conceivable that increased TGF release in the alveoli of patients with spontaneous remission renders T cells inactive and inhibits blast formation. This phenomenon was most prominent in the BAL and in the TBB, but in some cases it could even be observed in the PB.

There are several advantages in the usage of T cell clones for TCR analysis as opposed to the frequently applied methods of flowcytometry or PCR of the tissue. One important aim of this study was to compare the cell composition of the interstitium with that of the BAL. The number of T cells isolated from TBB of the interstitial tissue is not sufficient for flow cytometry. Histological methods for analyzing TCR-V-usage fail because of the high number of V families needing to be analyzed. Another advantage is that the established clones are available for other investigations. Finally, PCR methods do not allow the attribution of the disparities in the TCR-family usage to CD4⁺ or CD8⁺ T cell subpopulations.

However, the cloning method has two disadvantages. First, although we have cloned a large number of T cell clones, they represent only a small fraction of the entire T cells in the BAL. Additionally, the cloning procedure bears the risk of selection of T cells due to different proportions of clonable cells in the different compartments. Nevertheless, cloning T cells from BAL and PB we got comparable results as reported in the literature for these compartments (1, 7, 8, 19). From this we conclude that the cloning procedure in TBB reveals a representative distribution of the TCR-families in the lung interstitium.

Although the CD4/CD8 ratios of the clones are higher than the CD4/CD8 ratios of the original cell populations there is a positive correlation between these two parameters showing that the cloning procedure did not alter, but rather amplified, the CD4/CD8 ratios of the pre-cloning population. This is in accordance with Becker and colleagues, who also found an increased CD4/CD8 ratio after cloning compared with the ratios of the original cell population (22).

The CD4/CD8 ratio of the blood and the BAL from control subjects is normally about 2, and in the interstitium around 1 (23). The CD4/CD8 ratio of the clones derived from the blood is therefore in the normal range, whereas the CD4/ CD8 ratios of the clones from TBB and BAL are increased. This increase is more prominent in the BAL and less so in the interstitium. The striking correlation of the increase of the BAL and TBB CD4/CD8 ratios indicates the interrelationship of these compartments, whereas the lack of such correlations of the CD4/CD8 ratios of the blood with the BAL or TBB underlines the compartmentalization of sarcoidosis to the lung. Despite the fact that sarcoidosis is an interstitial lung disease, BAL is widely used to confirm the diagnosis or to assess the activity of the disease. However, it is not obvious that alveoli and interstitium are similar in their cell composition because they are separated by epithelial cells and the basal membrane. Nevertheless, several authors have found that BAL reflects the process taking place in the lung interstitium (14), which is in accordance with our results regarding the CD4/CD8 ratios.

A quite different picture emerges in the usage of TCR V5 and V8. Although the percentages of both families are significantly or nearly significantly increased in all three compartments, the significant difference of BAL and PB or TBB and the similarity of TBB and PB show the close relationship between these two compartments.

The spectrum of the TCR-family repertoire in the compartments alveoli and PB is under current investigation; however, nothing is known about the relationship between interstium and the two other compartments. Recent publications indicate that the distribution of the TCR families in the BAL does not differ from the distribution in the blood (1, 24). From this, expansion of certain TCR-families in the BAL or TBB should be regarded as antigen induced.

In sarcoidosis, the inhalation of an antigen may stimulate preferentially CD4⁺ cells of certain V families, e.g., V5 and V8 in the alveoli, the compartment of its entry. Such a selection, especially of V5, was demonstrated by Klein and coworkers in the intradermal lesions of the Kveim-Siltzbach reaction (10). This lead to an increase of the CD4/CD8 ratio and the absolute and relative number of the activated V families by proliferation and chemotaxis. However, in the interstitium and the peripheral blood, where either no antigen is present or is encapsulated in the granulomata, no proliferation or selection of V families occurs and their relative proportions are left unchanged. It is possible that in sarcoid patients the increase of the CD4/CD8 ratio in the interstitium is an unspecific feature of these inflammatory processes in the closely related alveoli and might be driven by soluble mediators released in the alveolar inflammatory process. One important mediator of T cell chemotaxis is the newly discovered cytokine IL-16, which induced a strong chemotactic signal in CD4⁺ T cells but not in CD8⁺ T-cells (25). The moderate increase of the interstitial CD4/CD8 ratio may reflect a transient state in the interstitium due to the migration of T cells between blood and the alveoli, and the withholding of the specific V families. Our findings of a relative separation of lung alveoli and interstitium are supported by a recent report by Nagata and coworkers, who could not correlate alveolar septal inflammation with BAL lymphocytosis in sarcoidosis (26).

In humans, a relative stability of the peripheral V repertoire over time in a single individual and substantial variation between different individuals in the population was reported (27). In healthy volunteers, it was shown that the CD4 V repertoire in blood differs significantly from the CD8 repertoire in a number of important ways. CD8 T cell repertoire of V2 and V3 is shown to be skewed, with an excess of individuals having higher values than consistent with a normal distribution (28). Surprisingly, a high degree of oligoclonality in CD8⁺ cells of normal subjects and persistence of TCR clonality over months were shown by Hingorani and associates (29). In our patients, many changes of the V-percentage could only be demonstrated in the CD4⁺ subpopulation, which may be expected due to the increase in the CD4/CD8 ratio in sarcoidosis. Saltini and colleagues demonstrated oligoclonality in the CD4⁺ subpopulation of patients with chronic beryllium disease, a disorder that is also accompanied by an increased CD4/ CD8 ratio and mimics many symptoms of sarcoidosis (30), and Gulwani-Akolkar detected the same pattern of selective expansion in the colon of patients with Crohn's disease (31). Conversely, in HIV infection, where the CD4/CD8 ratio substantially decreases, oligoclonal expansions of certain V families were demonstrated in the CD8⁺ subpopulation (32). In diseases without changes of the CD4/CD8 ratio e.g, Wegener's granulomatosis or polyarteritis nodosa, the skewed usage of V receptors could not be attributed to one subpopulation (33). Interestingly, in the blood of the HP patient, the V5 family could only be detected in the CD8⁺ subpopulation, whereas the increase of the same family in the BAL affected only the CD4⁺ subpopulation. In a recent publication, Wahlström and coworkers demonstrated that in T cell expansions in HP patients affect predominantly the CD8⁺ subpopulation which is in accordance to this case (34). In the TBB a normal percentage of the V5 family was detected.

The nature of the assumed "sarcoid antigen" is still under investigation. The heterogeneity of the increased V families described in the literature (9, 10, 19) make it unlikely that this antigen is a superantigen. Whether the "sarcoid antigen" stimulates certain V families to proliferate leading to an over-representation of these families or whether the skewing is based on increased recruitment of the cells remains unclear and cannot be clarified by these data. The interindividual variations of the T cell biases may be related to the genetic background of the patients, e.g., the MHC class II molecules, which have strong influence on the TCR composition of the T cells stimulated by a given antigen (35).

In conclusion, our data demonstrate that the immune reaction in sarcoidosis are strongly compartmentalized and that, although there are similarities between interstitium and alveoli, both compartments differ in the composition of their T cell subpopulations.